Using Propensity Score Matching in Program Evaluation: Lessons Learned in a Statewide Evaluation of a College-Prep Program in Hawai'i

• In Event: 31.092 - Poster Session 5
  In Poster Session: 31.092-1 - Methodological Approaches to Program Evaluation in the Schools

Abstract

This session will describe lessons learned in conducting a rigorous evaluation of a statewide college readiness program. While randomized controlled trials still remain the “gold standard” for identifying the causal impact of educational programs on student outcomes, the option to conduct them is often very slim. In contrast, secondary analysis of statewide data using propensity score matching is both feasible and improves the rigor of an evaluation. We find that matching, combined with covariate adjustment, is more effective in removing selection bias than either method used alone. Overall, our study provides solid empirical evidence for the policymakers’ decision to scale up the college-prep program, which began as a pilot, to a statewide program that reaches all high school students.

Authors

• Malkeet Singh, Education Northwest

• Michelle Hodara, Education Northwest

• Changhua Wang, Northwest Regional Educational Laboratory